Boinc 7.0.24/Ubuntu computation errors


log in

Advanced search

Message boards : Number crunching : Boinc 7.0.24/Ubuntu computation errors

Previous · 1 · 2 · 3 · 4 · 5 · Next
Author Message
Profile Karsten Vinding
Volunteer tester
Send message
Joined: 18 May 99
Posts: 140
Credit: 16,737,446
RAC: 1,165
Denmark
Message 1226238 - Posted: 2 May 2012, 16:43:51 UTC - in response to Message 1226223.

Oh well, I give up.

Even though I followed the instructions and installed Boincs own 7.0.26 and not 7.0.24 from repository, I still get this:

http://setiathome.berkeley.edu/result.php?resultid=2418178646

http://setiathome.berkeley.edu/result.php?resultid=2417963293

Both AP and MB

Computer is this one, many many failed tasks...

http://setiathome.berkeley.edu/results.php?hostid=6637638

Luckily I got it stopped before it ruined all my allocated WU's.

I'll wait for this to get fixed properly!

But thanks for trying to help xclusive :)
____________

JohnDKProject donor
Volunteer tester
Avatar
Send message
Joined: 28 May 00
Posts: 873
Credit: 48,314,128
RAC: 29,557
Denmark
Message 1226245 - Posted: 2 May 2012, 16:57:14 UTC - in response to Message 1226238.

Oh well, I give up.

Even though I followed the instructions and installed Boincs own 7.0.26 and not 7.0.24 from repository, I still get this:

http://setiathome.berkeley.edu/result.php?resultid=2418178646

http://setiathome.berkeley.edu/result.php?resultid=2417963293

Both AP and MB

Computer is this one, many many failed tasks...

http://setiathome.berkeley.edu/results.php?hostid=6637638

Luckily I got it stopped before it ruined all my allocated WU's.

I'll wait for this to get fixed properly!

But thanks for trying to help xclusive :)

But I see the error WUs reports using 7.0.24, so you must have BOINC 7.0.24 installed. Maybe doing what xclusive585 said will help, it deletes 7.0.24

Do "sudo apt-get remove boinc-manager"
and "sudo apt-get remove boinc-client"

Profile Ex
Volunteer moderator
Volunteer tester
Avatar
Send message
Joined: 12 Mar 12
Posts: 2895
Credit: 1,797,699
RAC: 396
United States
Message 1226246 - Posted: 2 May 2012, 16:59:38 UTC
Last modified: 2 May 2012, 17:07:56 UTC

Glad I helped you get it running, but it seems like the w-u's are failing for some other reason than mine were. blah. I'm sorry.

This is why I'm pushing the *buntu/Debian package folks for a proper fix.

This is unacceptable. I have a running Boinc, that's about it. They need to work this out, or many causal Linux folks will say screw this.
____________
-Dave #2

3.2.0-33

Profile Ex
Volunteer moderator
Volunteer tester
Avatar
Send message
Joined: 12 Mar 12
Posts: 2895
Credit: 1,797,699
RAC: 396
United States
Message 1226247 - Posted: 2 May 2012, 17:01:05 UTC

There should be a way to install the boinc package from the 10.04 repository into 12.04.... This would give you a properly working install, just a little older...

I don't have time now, but perhaps later I'll look into how to do that.
____________
-Dave #2

3.2.0-33

Profile Karsten Vinding
Volunteer tester
Send message
Joined: 18 May 99
Posts: 140
Credit: 16,737,446
RAC: 1,165
Denmark
Message 1226249 - Posted: 2 May 2012, 17:05:54 UTC - in response to Message 1226247.

John:

I did.

Do "sudo apt-get remove boinc-manager"
and "sudo apt-get remove boinc-client"

And the original repository install was uninstalled via the repository.

But I see your point about it being 7.0.24, but it should not be.

I will try one last time to install 7.0.26 on top of all this. If this fails I'm going to stop for now.

____________

Profile Ex
Volunteer moderator
Volunteer tester
Avatar
Send message
Joined: 12 Mar 12
Posts: 2895
Credit: 1,797,699
RAC: 396
United States
Message 1226252 - Posted: 2 May 2012, 17:12:36 UTC - in response to Message 1226249.
Last modified: 2 May 2012, 17:37:07 UTC

John:

I did.

Do "sudo apt-get remove boinc-manager"
and "sudo apt-get remove boinc-client"

And the original repository install was uninstalled via the repository.

But I see your point about it being 7.0.24, but it should not be.

I will try one last time to install 7.0.26 on top of all this. If this fails I'm going to stop for now.


There must be left over files in there that the manager is finding. Unfortunately when I asked in the recent past what/where files may be left-over I never really got an answer. I'd say delete any boinc or .boinc in /var/lib/, but that will also probably delete the lib files that boinc depends on (the ones that caused the error earlier) so is probably not a good option.

Blah later, I'll post detailed instructions on installing boinc in 12.04, from the 10.04(or similar) repository. This will be a "clean" method, to overcome the mess of issues from the current repo packages and manual install methods.
Unless someone else wants to look into that and post it. (Martin) :-)

(I would think you could speicfy a 10.04 repository to apt, in the same line you use to install boinc from apt. I.E: "sudo apt-get install boinc <from><ubuntu 10.04's repository>")(and before we even do this, we need to purge all of boinc's existing files from our computers. using "apt-get purge boinc", "apt-get autoremove"(<-DON'T do this if you are an advanced user with manually installed wares), deleting boinc files from /var/lib/, and manually deleting ALL files we've added to other locations.)
____________
-Dave #2

3.2.0-33

Profile Karsten Vinding
Volunteer tester
Send message
Joined: 18 May 99
Posts: 140
Credit: 16,737,446
RAC: 1,165
Denmark
Message 1226264 - Posted: 2 May 2012, 17:40:51 UTC - in response to Message 1226252.

I actually got it to run now.

Reporting 7.0.26 in Boinc Manager.

Its crunching away on two AP WU's and DL'ing more.

And I have merged the computer with the old install, so its downloading the rest of the old WU's.

Now on to intalling optimized apps, and making it start on bootup. Hope I dont mess it up...
____________

Profile Ex
Volunteer moderator
Volunteer tester
Avatar
Send message
Joined: 12 Mar 12
Posts: 2895
Credit: 1,797,699
RAC: 396
United States
Message 1226266 - Posted: 2 May 2012, 17:48:26 UTC
Last modified: 2 May 2012, 17:53:59 UTC

Good luck friend. one thing at a time. this way if you break it you'll know how to change it right back.

(and optimized aps is not anything I've done yet, I was waiting to get my server up to 12.04 first, (and that broke boinc), and now I'm gonna wait to have a clean repository install of 7.something.thatworks, but I know even running at 50% that would greatly increase my output)
____________
-Dave #2

3.2.0-33

Profile Karsten Vinding
Volunteer tester
Send message
Joined: 18 May 99
Posts: 140
Credit: 16,737,446
RAC: 1,165
Denmark
Message 1226267 - Posted: 2 May 2012, 17:59:14 UTC - in response to Message 1226266.
Last modified: 2 May 2012, 18:02:50 UTC

Do I have a problem in having installed BOINC as root?

It doesn't seem to want to run without me doing sudo -s, before launching it.

It works when done that way, but I cant see its task in the system monitor.

Got the optimized apps running, so that works.

When the repository version is reported working, I will install that, and use it instead of this hack-job install :)
____________

Profile Ex
Volunteer moderator
Volunteer tester
Avatar
Send message
Joined: 12 Mar 12
Posts: 2895
Credit: 1,797,699
RAC: 396
United States
Message 1226276 - Posted: 2 May 2012, 18:20:47 UTC
Last modified: 2 May 2012, 18:22:28 UTC

IT's not really a problem if it's running as root.

It works when done that way, but I cant see its task in the system monitor.

I'm not sure about desktop ubuntus, but I'm assuming the tasks aren't showing in your monitor because they are running as root, and your viewing the monitor as a user, there may be an option to "see all tasks" like in windows but I really don't know.

Try opening a terminal and running "sudo top", or even just "top", that should show you ALL running tasks.

(Also, if your boinc is runnning as root, and you use my crontab hack to run on start, you will need to put the command in the "root" cron by switching to root with sudo su before making the cron file.)
____________
-Dave #2

3.2.0-33

Profile Khangollo
Avatar
Send message
Joined: 1 Aug 00
Posts: 245
Credit: 36,410,524
RAC: 0
Slovenia
Message 1226281 - Posted: 2 May 2012, 18:31:53 UTC
Last modified: 2 May 2012, 18:36:59 UTC

ps -auxf

Will give you a complete process tree. Use Shift+Pg Up/Pg Down to scroll up and down the console.
Or, if you want to filter it to just some processes with "boinc" name in it,
ps -auxf | grep boinc

(or just ps -axf if the output is too wide; will not show the user accounts under which processes run)
____________

Profile Karsten Vinding
Volunteer tester
Send message
Joined: 18 May 99
Posts: 140
Credit: 16,737,446
RAC: 1,165
Denmark
Message 1226288 - Posted: 2 May 2012, 18:58:02 UTC - in response to Message 1226281.
Last modified: 2 May 2012, 19:11:14 UTC

Thanks.

"ps axf | grep AK" shows that the optimized clients are running on all 4 cores.

karsten@Ubuntu-server:~$ ps axf | grep AK
14085 pts/0 RNl+ 57:24 | \_ ../../projects/setiathome.berkeley.edu/AK_V8_linux64_sse3
14086 pts/0 RNl+ 22:55 | \_ ../../projects/setiathome.berkeley.edu/AK_V8_linux64_sse3
30593 pts/0 RNl+ 22:18 | \_ ../../projects/setiathome.berkeley.edu/AK_V8_linux64_sse3
31895 pts/0 RNl+ 19:41 | \_ ../../projects/setiathome.berkeley.edu/AK_V8_linux64_sse3

So all seems to be working. Work is also being reported back and even validated.

I also tried adding the crontab as root, will see if it works on next reboot.
____________

Profile Ex
Volunteer moderator
Volunteer tester
Avatar
Send message
Joined: 12 Mar 12
Posts: 2895
Credit: 1,797,699
RAC: 396
United States
Message 1226526 - Posted: 3 May 2012, 4:45:20 UTC

^[Off topic, but this^ is the beginning of a Linux junkie. It's how it all starts :-)]

Profile Ex
Volunteer moderator
Volunteer tester
Avatar
Send message
Joined: 12 Mar 12
Posts: 2895
Credit: 1,797,699
RAC: 396
United States
Message 1226767 - Posted: 3 May 2012, 18:01:59 UTC
Last modified: 3 May 2012, 18:07:52 UTC

Looking good...

So tonight (or as soon as my AP units all finish) I will be testing a new repository package of 7.0.26. :-)

Compiled specifically as a solution to our issue.

If it works that package will be pushed into the Ubuntu repository for 12.04!

I will update if it works, and update again if/when it's put in the repository.



:-D :-D
____________
-Dave #2

3.2.0-33

Richard HaselgroveProject donor
Volunteer tester
Send message
Joined: 4 Jul 99
Posts: 8760
Credit: 52,710,784
RAC: 24,247
United Kingdom
Message 1226778 - Posted: 3 May 2012, 18:34:03 UTC - in response to Message 1226767.

I don't want to be a party-pooper, but the BOINC trunk code is up to v7.0.27 now, and the Linux build - the last of the three major platforms to be built - was put online about an hour ago.

The bugfixes in .27 are less dramatic than the problems in .24 that started this thread, but they are pretty important, just the same. I wouldn't presume to advise you on package maintenance - there are arguments both ways, both for incremental fixes and for holding back for a single 'best of breed' one-time upgrade - but I just thought I'd better warn you that .26 isn't the end of the line.

Profile Ex
Volunteer moderator
Volunteer tester
Avatar
Send message
Joined: 12 Mar 12
Posts: 2895
Credit: 1,797,699
RAC: 396
United States
Message 1226781 - Posted: 3 May 2012, 18:39:36 UTC
Last modified: 3 May 2012, 19:33:39 UTC

Yes, but I'm speaking only about the Ubuntu repository versions which have had an issue since 7.0.24.

[The Debian and Ubuntu package maintainers are aware that their compiles have an error affecting s@h and possibly others. They are trying to figure out what it is, and I think in time they will. In the meantime the Ubuntu package maintainer has been nice enough to compile one for us that doesn't have the debian patches that are assumed to be breaking it. As I said once it's tested and works they will be using it in the standard Ubuntu 12.04 repo, as a band-aid until they really get their issue nailed]

They expect 7.0.27 wont have this issue when they compile it for the repository, the Debian guys are looking into it. One thing at a time. Even 7.0.26 is too new for the standard repository IMO, but if this specific compile has been built in a fashion that actually will work, than that's the package they should use in the standard repository for now...

In the meantime we (Ubuntu users) will at least have a working version that the causal user can install with a package manager. [People use Ubuntu for simplicity, it defeats the purpose if they can't use a package manager to get crunching s@h with Boinc]

When they get the repository package bumped to 7.0.27 (or something else) is another issue.

But any working package in the Ubuntu repository will be better than the one that doesn't work. :-)
____________
-Dave #2

3.2.0-33

Profile Ex
Volunteer moderator
Volunteer tester
Avatar
Send message
Joined: 12 Mar 12
Posts: 2895
Credit: 1,797,699
RAC: 396
United States
Message 1227624 - Posted: 5 May 2012, 5:27:10 UTC

So we know that the repository package is being broken (for S@H) due to a Debian patch used while compiling the Debian/Ubuntu repository package of Boinc 7.0.24.

Through process of elimination we will figure out which patch needs to be fixed.

A fix is in the works, that's all I can offer for now.

And I apologize for the few errored or dropped work-units I have and will produce here and there. Sorry wingmen. I'm keeping it to an absolute minimum. Apparently as the one bringing their bug to their attention, I get the honor of testing for them. LOL

Steffen Moeller, who is the Debian package maintainer, is aware of the issue. He is working with us to figure out which Debian/Ubuntu patch needs to be fixed. Because of this future repository packages should be all set. :-)

Steffen had this to offer us:

"The bug is fixed by undoing some presumed improvements of mine. This needs some further investigation...
Maybe we should just fine those aliens quickly and have them fix it all for us...or eradicate man kind.

Cheers,

Steffen"


So, just hang out for a fix, we're making it happen. Testing one build at a time.
____________
-Dave #2

3.2.0-33

Profile Ex
Volunteer moderator
Volunteer tester
Avatar
Send message
Joined: 12 Mar 12
Posts: 2895
Credit: 1,797,699
RAC: 396
United States
Message 1228519 - Posted: 7 May 2012, 1:45:22 UTC

We're pretty sure we've narrowed down the repository bug to one specific patch! Luckily it's not a patch crucial to anything, as far as the Debian package maintainer believes.

Hopefully tomorrow the build will be put together and tested.

If it's a good build I think it will be pushed into the Ubuntu repository. This would be version 7.0.26 afaik.

7.0.27 is having issues being compiled at all by the Ubuntu folks, something is awry. So if this working version 7.0.26 ends up in the Ubuntu repository, we probably wouldn't see another version change in any Debian/Ubuntu repository until .28 at the soonest.

I'll update tomorrow or the day after if/when they put a working package in the Ubuntu repository.
____________
-Dave #2

3.2.0-33

Profile Ex
Volunteer moderator
Volunteer tester
Avatar
Send message
Joined: 12 Mar 12
Posts: 2895
Credit: 1,797,699
RAC: 396
United States
Message 1228707 - Posted: 7 May 2012, 15:31:31 UTC
Last modified: 7 May 2012, 15:32:19 UTC

:-D

The Debian/Ubuntu repository "instant computation error" bug is now known to be caused by "MoreInformativeAppStartFailure.patch"

I think the name of this patch is ironic considering the bug that it caused... (Perhaps they should have just named it "AppStartFailure.Patch")

There was some talk between the package maintainers about trying a 7.0.27 build without this patch, and testing that for the repository. We'll see what happens.

As I said were getting it figured out, a fix will be in the repository shortly.

:-)
____________
-Dave #2

3.2.0-33

Profile Ageless
Avatar
Send message
Joined: 9 Jun 99
Posts: 12395
Credit: 2,666,323
RAC: 958
Netherlands
Message 1228725 - Posted: 7 May 2012, 17:18:12 UTC
Last modified: 7 May 2012, 17:33:15 UTC

Edit: Link was already given earlier...
____________
Jord

Fighting for the correct use of the apostrophe, together with Weird Al Yankovic

Previous · 1 · 2 · 3 · 4 · 5 · Next

Message boards : Number crunching : Boinc 7.0.24/Ubuntu computation errors

Copyright © 2014 University of California